Segments, syllables and speech tempo perception

Abstract

Studies of speech tempo commonly use syllable or segment rate as a proxy measure for perceived tempo. In languages whose phonologies allow substantial syllable complexity these measures can produce figures on quite different scales. Listeners’ sensitivity to syllable rate has been demonstrated in multiple studies in which listeners judge the rhythm or tempo of spoken utterances, although these studies do not control for segment rate. Evidence for listeners’ sensitivity to segment rate is much rarer. We report two experiments aimed at clarifying the contributions of syllable and segment rate to English listeners’ tempo judgements. In the first experiment, we manipulate syllable rate in utterance pairs that are constant in segment rate; in the second, we keep syllable rate constant and manipulate segment rate. Listeners decide for each pair which utterance sounds faster. Our results suggest that syllable rate differences are perceived as tempo differences even if segment rate is constant, while differences in segment rate that do not correspond to differences in syllable rate have little impact on perceived speech tempo in English.